Morning Brief

Despite the popularity of Claude Code, Cursor, GitHub Copilot, and OpenAI Codex, developers have plenty of complaints about AI coding tools. So researchers affiliated with York University and the University of Calgary in Canada decided to sift through developers' concerns about LLM-based integrated development environments (LIDEs) by analyzing Reddit discussions for common themes. Their findings suggest that the builders of such tools failed to prioritize security and privacy, leaving developers to defend themselves. Gias Uddin, associate professor at York University and a co-author of the…

A newly disclosed Linux kernel vulnerability, dubbed SCTPhantom and tracked as CVE-2026-64564, allows attackers to escalate from unprivileged local access to full root and even escape containers to compromise the underlying host. The flaw is a use-after-free bug in the kernel’s SCTP Dynamic Address Reconfiguration feature, and remarkably, its root cause traces back to code introduced in Linux 2.6.25 in December 2007, making it nearly 18 years old before discovery. The vulnerability lives in how the kernel handles ASCONF chunks, a mechanism defined in RFC 5061 that lets SCTP associations add…

A Linux kernel vulnerability, tracked as CVE-2026-64561 and named Zapscape, could allow attackers to escape a KVM virtual machine and take control of its underlying Linux host with root privileges. The issue affects KVM/x86, a virtualization technology that separates guest systems from the physical server. The flaw is especially serious for cloud providers and enterprises that run untrusted workloads. Zapscape was discovered by security researcher Hyunwoo Kim, known as V4bel. It exists in KVM’s shadow memory management unit, or shadow MMU. It manages memory translations when nested…

PortSwigger says HTTP Terminator, an artificial intelligence (AI)-assisted research system built by James Kettle, generated and proved new HTTP desynchronization techniques after exploring 30,000 candidate desync vectors. PortSwigger said a separate human-guided discovery cascade also exposed a zero-day in Apache Traffic Server. Kettle said HTTP Terminator tested 30,000 websites where

ClickFix-style attacks are being used to deliver a Go-based malware capable of stealing cryptocurrency assets, as well as browser-stored passwords, Apple iCloud Keychain data, and cached credentials. The macOS-focused infection chain is designed to deliver a shell script that profiles the host and then fetches a macOS malware payload that's compatible with the computer's CPU architecture. "

A cluster of nearly 800 malicious packages has been published to the npm registry as part of a new campaign designed to deliver cross-platform malware targeting Windows, Mac, and Linux systems. "These packages appear to use AI slop squatted, or randomly generated typo-squatting package names, but all of them deliver a powerful RAT and infostealer payload," OpenSourceMalware researcher Paul

A newly demonstrated technique shows how malware in a compromised Windows user session can abuse Windows Hello for Business (WHFB) cryptographic keys to authenticate to Microsoft Entra ID, enabling attackers to gain cloud access without the victim’s password, PIN, or biometric data. Windows Hello for Business is designed as a passwordless authentication system. It normally stores a user’s private key in the device’s Trusted Platform Module, or TPM, making the key difficult to export or steal. Users unlock access to that key with a PIN, fingerprint, facial recognition, or another local…

Attacker-controlled instructions can make Atlassian's Rovo assistant collect Jira or Confluence data that a signed-in user can access, then send it to an outside server. Two security firms found that behavior independently, by different routes. Only one of those routes is confirmed closed. PromptArmor, an AI security firm, hid the instructions in content Rovo reads. It said an uploaded file was